Topics

IBMTTS and eloquence driver comments.


DaVid
 

Hi all.

I'd like to comment about eloquence and IBMTTS legality. Yes, I'm
reliving a very past commented topic, I know.
I was the first developer of the driver, originally called ibmeci or
eci, I can't remember. At time of developing it, I was a high school
teenager unconcerned of licence problems.

But develop for IBMTTS is not ilegal I think. Voxin and TTSynth let it
for theyr products, but the user must have to buy a personal licence.
E.G. to oralux, it costs just $5 per language.
So the driver IBMTTS should be legal.
Although use it with eloquence libraries would be ilegal and I as a
developer can't control it, it's just life coincidences that eloquence
and IBMTTS share the same api.
OK, voxin or TTSynth doesn't work on windows, but it could be solved
with wsl. I had good results running voxin on wsl, and I could write a
fakade eci library on windows to communicate with voxin in wsl. Oralux
has in their github an eci fakade example.
I just want show more evidence about legality of an IBMTTS driver.
NVDA users could buy voxin and use it or any library with the same
api.

Some weeks ago I loaded a IBMTTS driver to github as a NVDA addon, it
adds many new features and compatibility with python 3, audio ducking
improvements, easier dialog to configure the path of IBM tts library
files, ETC. Also contributions from many of you, but as a ilegal
driver no one wrote his name here.

I'd like to receive contributions here to avoid multiple driver distributions.
So I want to ask you. If this driver is based on voxin documentation
(And no one can prove otherwise) could it be legal?
Do you consider useful to make voxin compatible with NVDA?

I know about CF driver. But that driver is expensive, outdated, with
many performance problems, requires internet connection, ETC. And CF
doesn't seem to want to solve any of this.
Could be useful to use a IBM tts voice like voxin with NVDA for just $5?

Regards,
David.


Patrick Perdue <borrisinabox@...>
 

Hi:


As far as I'm concerned, the Code Factory package is essentially unusable. At best, it's glitchy and unstable, and, yes, that licensing scheme sure is annoying. I have computers I keep offline for specific purposes, and that isn't compatible with such operations.

This is why I never bought it.


Honestly, I'm no fan of Eloquence, but it's what I use, because my hearing has gotten to the point to where I can't really deal with eSpeak anymore, which I used for several years when I first started using NVDA.

The one core voices don't do it for me either, as they break in strange ways, even at slow speed.

So, I'm shamelessly using the not so secret Eloquence driver just so I can get work done.

I'd probably pay for a personal Voxen license to insure future compatibility or whatever.

On 4/23/2019 8:52 PM, DaVid wrote:
Hi all.

I'd like to comment about eloquence and IBMTTS legality. Yes, I'm
reliving a very past commented topic, I know.
I was the first developer of the driver, originally called ibmeci or
eci, I can't remember. At time of developing it, I was a high school
teenager unconcerned of licence problems.

But develop for IBMTTS is not ilegal I think. Voxin and TTSynth let it
for theyr products, but the user must have to buy a personal licence.
E.G. to oralux, it costs just $5 per language.
So the driver IBMTTS should be legal.
Although use it with eloquence libraries would be ilegal and I as a
developer can't control it, it's just life coincidences that eloquence
and IBMTTS share the same api.
OK, voxin or TTSynth doesn't work on windows, but it could be solved
with wsl. I had good results running voxin on wsl, and I could write a
fakade eci library on windows to communicate with voxin in wsl. Oralux
has in their github an eci fakade example.
I just want show more evidence about legality of an IBMTTS driver.
NVDA users could buy voxin and use it or any library with the same
api.

Some weeks ago I loaded a IBMTTS driver to github as a NVDA addon, it
adds many new features and compatibility with python 3, audio ducking
improvements, easier dialog to configure the path of IBM tts library
files, ETC. Also contributions from many of you, but as a ilegal
driver no one wrote his name here.

I'd like to receive contributions here to avoid multiple driver distributions.
So I want to ask you. If this driver is based on voxin documentation
(And no one can prove otherwise) could it be legal?
Do you consider useful to make voxin compatible with NVDA?

I know about CF driver. But that driver is expensive, outdated, with
many performance problems, requires internet connection, ETC. And CF
doesn't seem to want to solve any of this.
Could be useful to use a IBM tts voice like voxin with NVDA for just $5?

Regards,
David.


 

Well code factory hold the rights for eloquence so we have to put up with it.

Then again do you need to use eloquence.

There are the 1 core voices, and vocaliser.

If you really, really need for windows eloquence or a similar, flite will work, its free, and crappy sounding its used on linux to.

Over that, voxin is only for linux but its cheap enough.

Talk to the devs and see what it would take to make it for nvda and or sapi.

I mean if its only 30 bucks an os I will buy it gladly.

On 24/04/2019 12:52 PM, DaVid wrote:
Hi all.

I'd like to comment about eloquence and IBMTTS legality. Yes, I'm
reliving a very past commented topic, I know.
I was the first developer of the driver, originally called ibmeci or
eci, I can't remember. At time of developing it, I was a high school
teenager unconcerned of licence problems.

But develop for IBMTTS is not ilegal I think. Voxin and TTSynth let it
for theyr products, but the user must have to buy a personal licence.
E.G. to oralux, it costs just $5 per language.
So the driver IBMTTS should be legal.
Although use it with eloquence libraries would be ilegal and I as a
developer can't control it, it's just life coincidences that eloquence
and IBMTTS share the same api.
OK, voxin or TTSynth doesn't work on windows, but it could be solved
with wsl. I had good results running voxin on wsl, and I could write a
fakade eci library on windows to communicate with voxin in wsl. Oralux
has in their github an eci fakade example.
I just want show more evidence about legality of an IBMTTS driver.
NVDA users could buy voxin and use it or any library with the same
api.

Some weeks ago I loaded a IBMTTS driver to github as a NVDA addon, it
adds many new features and compatibility with python 3, audio ducking
improvements, easier dialog to configure the path of IBM tts library
files, ETC. Also contributions from many of you, but as a ilegal
driver no one wrote his name here.

I'd like to receive contributions here to avoid multiple driver distributions.
So I want to ask you. If this driver is based on voxin documentation
(And no one can prove otherwise) could it be legal?
Do you consider useful to make voxin compatible with NVDA?

I know about CF driver. But that driver is expensive, outdated, with
many performance problems, requires internet connection, ETC. And CF
doesn't seem to want to solve any of this.
Could be useful to use a IBM tts voice like voxin with NVDA for just $5?

Regards,
David.


.


zvonimir stanečić, 9a5dsz
 

Hi to all,
It's not true that the codefactory driver i unnusable.
Even it works fine with the new NVDA release.

-----Original Message-----
From: nvda-addons@nvda-addons.groups.io <nvda-addons@nvda-addons.groups.io> On Behalf Of Patrick Perdue
Sent: Wednesday, April 24, 2019 4:41 AM
To: nvda-addons@nvda-addons.groups.io
Subject: Re: [nvda-addons] IBMTTS and eloquence driver comments.

Hi:


As far as I'm concerned, the Code Factory package is essentially unusable. At best, it's glitchy and unstable, and, yes, that licensing scheme sure is annoying. I have computers I keep offline for specific purposes, and that isn't compatible with such operations.

This is why I never bought it.


Honestly, I'm no fan of Eloquence, but it's what I use, because my hearing has gotten to the point to where I can't really deal with eSpeak anymore, which I used for several years when I first started using NVDA.

The one core voices don't do it for me either, as they break in strange ways, even at slow speed.

So, I'm shamelessly using the not so secret Eloquence driver just so I can get work done.

I'd probably pay for a personal Voxen license to insure future compatibility or whatever.


On 4/23/2019 8:52 PM, DaVid wrote:
Hi all.

I'd like to comment about eloquence and IBMTTS legality. Yes, I'm
reliving a very past commented topic, I know.
I was the first developer of the driver, originally called ibmeci or
eci, I can't remember. At time of developing it, I was a high school
teenager unconcerned of licence problems.

But develop for IBMTTS is not ilegal I think. Voxin and TTSynth let it
for theyr products, but the user must have to buy a personal licence.
E.G. to oralux, it costs just $5 per language.
So the driver IBMTTS should be legal.
Although use it with eloquence libraries would be ilegal and I as a
developer can't control it, it's just life coincidences that eloquence
and IBMTTS share the same api.
OK, voxin or TTSynth doesn't work on windows, but it could be solved
with wsl. I had good results running voxin on wsl, and I could write a
fakade eci library on windows to communicate with voxin in wsl. Oralux
has in their github an eci fakade example.
I just want show more evidence about legality of an IBMTTS driver.
NVDA users could buy voxin and use it or any library with the same
api.

Some weeks ago I loaded a IBMTTS driver to github as a NVDA addon, it
adds many new features and compatibility with python 3, audio ducking
improvements, easier dialog to configure the path of IBM tts library
files, ETC. Also contributions from many of you, but as a ilegal
driver no one wrote his name here.

I'd like to receive contributions here to avoid multiple driver distributions.
So I want to ask you. If this driver is based on voxin documentation
(And no one can prove otherwise) could it be legal?
Do you consider useful to make voxin compatible with NVDA?

I know about CF driver. But that driver is expensive, outdated, with
many performance problems, requires internet connection, ETC. And CF
doesn't seem to want to solve any of this.
Could be useful to use a IBM tts voice like voxin with NVDA for just $5?

Regards,
David.



DaVid
 

For Patrick: With my IBMTTS driver you can use eloquence or the older
viavoice if you want, but you could use voxin if I implement it from
wsl.
All the three share exactly the same api.

For Shaun: One core and vocalizer voices can't speak as fast as
Eloquence or Espeak.
I use eloquence at 90% and Espeak I don't remember, but I always need
to enable rate boost.
OK maybe vocalizer, but vocalizer is not very understantable at high rates.
Advanced blind users need to be fast in using computers, so we need
fast and understandable synthesizers.
Espeak could be good in english, but is not very good in spanish or
another languages.
I tried to recommend Espeak to new users, but no one liked it.
Bad point for NVDA, and I can't force users to use Espeak. I need to
attract users to this screen reader, so I am looking for a cheap
solution. Vocalizer or eloquence from CF aren't cheap solutions.

Voxin can't be compiled for windows because Oralux doesn't has the source code.
If we want to run voxin with NVDA, we need to use WSL or similar
system to run linux software. I wrote them, they told me that I can
implement a driver for my app, but the final user needs to buy a voxin
copy.
My app = A bridge controller for NVDA, from linux to windows.

I don't know about flite, I will look for a audio demo.

Regards,
David.

2019-04-23 21:15 GMT-06:00, Shaun Everiss <sm.everiss@gmail.com>:

Well code factory hold the rights for eloquence so we have to put up
with it.

Then again do you need to use eloquence.

There are the 1 core voices, and vocaliser.

If you really, really need for windows eloquence or a similar, flite
will work, its free, and crappy sounding its used on linux to.

Over that, voxin is only for linux but its cheap enough.

Talk to the devs and see what it would take to make it for nvda and or
sapi.

I mean if its only 30 bucks an os I will buy it gladly.



On 24/04/2019 12:52 PM, DaVid wrote:
Hi all.

I'd like to comment about eloquence and IBMTTS legality. Yes, I'm
reliving a very past commented topic, I know.
I was the first developer of the driver, originally called ibmeci or
eci, I can't remember. At time of developing it, I was a high school
teenager unconcerned of licence problems.

But develop for IBMTTS is not ilegal I think. Voxin and TTSynth let it
for theyr products, but the user must have to buy a personal licence.
E.G. to oralux, it costs just $5 per language.
So the driver IBMTTS should be legal.
Although use it with eloquence libraries would be ilegal and I as a
developer can't control it, it's just life coincidences that eloquence
and IBMTTS share the same api.
OK, voxin or TTSynth doesn't work on windows, but it could be solved
with wsl. I had good results running voxin on wsl, and I could write a
fakade eci library on windows to communicate with voxin in wsl. Oralux
has in their github an eci fakade example.
I just want show more evidence about legality of an IBMTTS driver.
NVDA users could buy voxin and use it or any library with the same
api.

Some weeks ago I loaded a IBMTTS driver to github as a NVDA addon, it
adds many new features and compatibility with python 3, audio ducking
improvements, easier dialog to configure the path of IBM tts library
files, ETC. Also contributions from many of you, but as a ilegal
driver no one wrote his name here.

I'd like to receive contributions here to avoid multiple driver
distributions.
So I want to ask you. If this driver is based on voxin documentation
(And no one can prove otherwise) could it be legal?
Do you consider useful to make voxin compatible with NVDA?

I know about CF driver. But that driver is expensive, outdated, with
many performance problems, requires internet connection, ETC. And CF
doesn't seem to want to solve any of this.
Could be useful to use a IBM tts voice like voxin with NVDA for just $5?

Regards,
David.



.



David Taylor
 

Hi,


This is it for me too, while it's not my hearing, I find my brain simply cannot process any other synth at the speed I need, and I too get thrown by the way the Code Factory addon breaks, which it does, very regularly. Modern voices, without exception, cause me many, many issues, not least being speed, terrible pronunciation issues and those awful pauses. Sadly, it's this simple. If that Eloquence driver breaks, and there is no replacement that sounds identical, I am forced back to Jaws. I have to say Speech Player was showing a lot of potential, and I think could have been made to work, but that development was dropped. I'm sorry to be so negative, but there literally is no other voice I can deal with on the computer


Windows One core voices are a great example. On any half decent set of speakers, something about them jars and gives me a headache, all the different languages have howling but different pronunciation issues, they are far to quiet, and even with the latest NVDA update, the punctuation pauses are far too long for doing any kind of business. Many of the voices also have too wide and unnatural intonations, making them very hard to listen to. I could literally use my old Braille 'n Speak faster than I can use most modern voices, although I doubt that would even work now we're used to Eloquence. Any replacement must have the precision of sound and pronunciation, without all the bass that's in many modern voices which are configured for phones, not for any decent sound system. I've bought both Acapella and Vocaliser drivers to try as well, and nothing comes close to Eloquence. The Amy voice from whoever it is is lovely, but too quiet and again, awful, awful  prununciations, so I can't even do anything quickly with that. So, my choices are, NVDA 19.1 forever, Jaws or somebody makes that addon work.


Cheers

Dave

On 24/04/2019 03:40, Patrick Perdue wrote:
Hi:


As far as I'm concerned, the Code Factory package is essentially unusable. At best, it's glitchy and unstable, and, yes, that licensing scheme sure is annoying. I have computers I keep offline for specific purposes, and that isn't compatible with such operations.

This is why I never bought it.


Honestly, I'm no fan of Eloquence, but it's what I use, because my hearing has gotten to the point to where I can't really deal with eSpeak anymore, which I used for several years when I first started using NVDA.

The one core voices don't do it for me either, as they break in strange ways, even at slow speed.

So, I'm shamelessly using the not so secret Eloquence driver just so I can get work done.

I'd probably pay for a personal Voxen license to insure future compatibility or whatever.


On 4/23/2019 8:52 PM, DaVid wrote:
Hi all.

I'd like to comment about eloquence and IBMTTS legality. Yes, I'm
reliving a very past commented topic, I know.
I was the first developer of the driver, originally called ibmeci or
eci, I can't remember. At time of developing it, I was a high school
teenager unconcerned of licence problems.

But develop for IBMTTS is not ilegal I think. Voxin and TTSynth let it
for theyr products, but the user must have to buy a personal licence.
E.G. to oralux, it costs just $5 per language.
So the driver IBMTTS should be legal.
Although use it with eloquence libraries would be ilegal and I as a
developer can't control it, it's just life coincidences that eloquence
and IBMTTS share the same api.
OK, voxin or TTSynth doesn't work on windows, but it could be solved
with wsl. I had good results running voxin on wsl, and I could write a
fakade eci library on windows to communicate with voxin in wsl. Oralux
has in their github an eci fakade example.
I just want show more evidence about legality of an IBMTTS driver.
NVDA users could buy voxin and use it or any library with the same
api.

Some weeks ago I loaded a IBMTTS driver to github as a NVDA addon, it
adds many new features and compatibility with python 3, audio ducking
improvements, easier dialog to configure the path of IBM tts library
files, ETC. Also contributions from many of you, but as a ilegal
driver no one wrote his name here.

I'd like to receive contributions here to avoid multiple driver distributions.
So I want to ask you. If this driver is based on voxin documentation
(And no one can prove otherwise) could it be legal?
Do you consider useful to make voxin compatible with NVDA?

I know about CF driver. But that driver is expensive, outdated, with
many performance problems, requires internet connection, ETC. And CF
doesn't seem to want to solve any of this.
Could be useful to use a IBM tts voice like voxin with NVDA for just $5?

Regards,
David.



José Manuel Delicado Alcolea
 

Hi David.

Although the WSL approach seems the optimal solution for licensing, I see many disadvantages for the medium user:

-Making the synthesizer work requires enabling WSL (Windows subsystem for Linux) support, installing Ubuntu or another distribution from the Microsoft Store, configuring it, installing Voxin, integrating it with Speech Dispatcher, configure Speech Dispatcher to listen on a tcp port and, finally, installing an NVDA add-on.

-WSL works like a container, similar to LXC and Docker. This means that the only running processes are those you run manually. There are no daemons, no services unless you start them, so Speech Dispatcher should be started manually before running NVDA.

-WSL requires Windows 10 Pro / Enterprise / Pro for workstations. Windows 10 home, 8, 8.1 and 7 are excluded.

In conclussion, only very advanced users could use Eloquence in this case. However, this story must finish as soon as possible. It has been opened for many years. We have discussed it in the spanish list many times, and I'd like to share the results and opened questions here:

-First of all: Code Factory Eloquence is too expensive for users who only want to use Eloquence in one computer. The add-on is great for those who will install the license in three computers and download Vocalizer voices, cheaper than Tiflotecnia's add-on.

-Code Factory add-on loads an additional process in order to work. This process includes an extra settings interface which allows ignoring and bypassing the NVDA synth settings ring configuration. In my opinion, this is a very bad practice, currently NVDA is flexible enough to allow many configurations for different situations.

-Illegal Eloquence add-on works much better than Code Factory add-on, is updated more frequently and is compatible with recent NVDA versions. It has many variants, David is trying to centralize all work in one repository.

-Illegal Eloquence is written in pure Python, except for Eloquence DLLS, of course. You can see the entire add-on source code, contribute with changes, create derived works and so on.

-Also, Tiflotecnia add-on is updated regularly and written in Python. I recommend purchasing and using it if you want high quality voices.

-First question: why is the illegal Eloquence illegal? We don't have a clear answer, but We believe the problem comes from the libraries distribution. Illegal Eloquence has not permission from Nuance to distribute the DLLS. Would it become legal if libraries are removed? We don't know.

-Who owns Eloquence currently? Nuance has the ownership, but Code Factory is the main distributor since 2014.

-Can I be a Code Factory reseller? Yes, but you have to purchase many pre-paid licenses before selling.

-Can I access the Eloquence SDK and create a legal version derived from the illegal add-on? Yes, but you have to pay €3500 to Code Factory in first place, and €18 per sold license.

-Can I become a Nuance distributor? Definitely, no. Seems Nuance and Code Factory have a special relationship, unreachable for others.

-Why is still Nuance selling a discontinued product? Good question, there is no answer. It should be open source, or at least free.

-Is the illegal Eloquence add-on source code legal? Many people believe this code has been written after reverse-engineering the Eloquence included with Jaws. This is not true. The add-on source code was written using the documentation provided in the IBM TTS SDK. This SDK was released as open source by IBM, but became closed after Nuance purchased the rights. So the real question is...

-What happens to a derived work if the original open source software becomes closed? Even reading the IBM TTS Sdk license, we don't know. David, could you please attach this document?

-And the final question: is David's add-on legal? If we answer the previous questions, this one will be answered too. Despite that, David, I think you might request a basic review.

Regards.


El 24/04/2019 a las 7:10, DaVid escribió:
For Patrick: With my IBMTTS driver you can use eloquence or the older
viavoice if you want, but you could use voxin if I implement it from
wsl.
All the three share exactly the same api.

For Shaun: One core and vocalizer voices can't speak as fast as
Eloquence or Espeak.
I use eloquence at 90% and Espeak I don't remember, but I always need
to enable rate boost.
OK maybe vocalizer, but vocalizer is not very understantable at high rates.
Advanced blind users need to be fast in using computers, so we need
fast and understandable synthesizers.
Espeak could be good in english, but is not very good in spanish or
another languages.
I tried to recommend Espeak to new users, but no one liked it.
Bad point for NVDA, and I can't force users to use Espeak. I need to
attract users to this screen reader, so I am looking for a cheap
solution. Vocalizer or eloquence from CF aren't cheap solutions.

Voxin can't be compiled for windows because Oralux doesn't has the source code.
If we want to run voxin with NVDA, we need to use WSL or similar
system to run linux software. I wrote them, they told me that I can
implement a driver for my app, but the final user needs to buy a voxin
copy.
My app = A bridge controller for NVDA, from linux to windows.

I don't know about flite, I will look for a audio demo.

Regards,
David.

2019-04-23 21:15 GMT-06:00, Shaun Everiss <sm.everiss@...>:
Well code factory hold the rights for eloquence so we have to put up
with it.

Then again do you need to use eloquence.

There are the 1 core voices, and vocaliser.

If you really, really need for windows eloquence or a similar, flite
will work, its free, and crappy sounding its used on linux to.

Over that, voxin is only for linux but its cheap enough.

Talk to the devs and see what it would take to make it for nvda and or
sapi.

I mean if its only 30 bucks an os I will buy it gladly.



On 24/04/2019 12:52 PM, DaVid wrote:
Hi all.

I'd like to comment about eloquence and IBMTTS legality. Yes, I'm
reliving a very past commented topic, I know.
I was the first developer of the driver, originally called ibmeci or
eci, I can't remember. At time of developing it, I was a high school
teenager unconcerned of licence problems.

But develop for IBMTTS is not ilegal I think. Voxin and TTSynth let it
for theyr products, but the user must have to buy a personal licence.
E.G. to oralux, it costs just $5 per language.
So the driver IBMTTS should be legal.
Although use it with eloquence libraries would be ilegal and I as a
developer can't control it, it's just life coincidences that eloquence
and IBMTTS share the same api.
OK, voxin or TTSynth doesn't work on windows, but it could be solved
with wsl. I had good results running voxin on wsl, and I could write a
fakade eci library on windows to communicate with voxin in wsl. Oralux
has in their github an eci fakade example.
I just want show more evidence about legality of an IBMTTS driver.
NVDA users could buy voxin and use it or any library with the same
api.

Some weeks ago I loaded a IBMTTS driver to github as a NVDA addon, it
adds many new features and compatibility with python 3, audio ducking
improvements, easier dialog to configure the path of IBM tts library
files, ETC. Also contributions from many of you, but as a ilegal
driver no one wrote his name here.

I'd like to receive contributions here to avoid multiple driver
distributions.
So I want to ask you. If this driver is based on voxin documentation
(And no one can prove otherwise) could it be legal?
Do you consider useful to make voxin compatible with NVDA?

I know about CF driver. But that driver is expensive, outdated, with
many performance problems, requires internet connection, ETC. And CF
doesn't seem to want to solve any of this.
Could be useful to use a IBM tts voice like voxin with NVDA for just $5?

Regards,
David.



.






--

José Manuel Delicado Alcolea
Administrador y editor en la web nvda.es
Twitter: @nvda_es
Certificado en el programa NVDA Expert 2019

Logo NVDA Certified Expert


Mohamed
 

No? Pretty sure WSL works on Windows 10 home as well.

On 4/24/2019 6:19 AM, José Manuel Delicado Alcolea via Groups.Io wrote:

Hi David.

Although the WSL approach seems the optimal solution for licensing, I see many disadvantages for the medium user:

-Making the synthesizer work requires enabling WSL (Windows subsystem for Linux) support, installing Ubuntu or another distribution from the Microsoft Store, configuring it, installing Voxin, integrating it with Speech Dispatcher, configure Speech Dispatcher to listen on a tcp port and, finally, installing an NVDA add-on.

-WSL works like a container, similar to LXC and Docker. This means that the only running processes are those you run manually. There are no daemons, no services unless you start them, so Speech Dispatcher should be started manually before running NVDA.

-WSL requires Windows 10 Pro / Enterprise / Pro for workstations. Windows 10 home, 8, 8.1 and 7 are excluded.

In conclussion, only very advanced users could use Eloquence in this case. However, this story must finish as soon as possible. It has been opened for many years. We have discussed it in the spanish list many times, and I'd like to share the results and opened questions here:

-First of all: Code Factory Eloquence is too expensive for users who only want to use Eloquence in one computer. The add-on is great for those who will install the license in three computers and download Vocalizer voices, cheaper than Tiflotecnia's add-on.

-Code Factory add-on loads an additional process in order to work. This process includes an extra settings interface which allows ignoring and bypassing the NVDA synth settings ring configuration. In my opinion, this is a very bad practice, currently NVDA is flexible enough to allow many configurations for different situations.

-Illegal Eloquence add-on works much better than Code Factory add-on, is updated more frequently and is compatible with recent NVDA versions. It has many variants, David is trying to centralize all work in one repository.

-Illegal Eloquence is written in pure Python, except for Eloquence DLLS, of course. You can see the entire add-on source code, contribute with changes, create derived works and so on.

-Also, Tiflotecnia add-on is updated regularly and written in Python. I recommend purchasing and using it if you want high quality voices.

-First question: why is the illegal Eloquence illegal? We don't have a clear answer, but We believe the problem comes from the libraries distribution. Illegal Eloquence has not permission from Nuance to distribute the DLLS. Would it become legal if libraries are removed? We don't know.

-Who owns Eloquence currently? Nuance has the ownership, but Code Factory is the main distributor since 2014.

-Can I be a Code Factory reseller? Yes, but you have to purchase many pre-paid licenses before selling.

-Can I access the Eloquence SDK and create a legal version derived from the illegal add-on? Yes, but you have to pay €3500 to Code Factory in first place, and €18 per sold license.

-Can I become a Nuance distributor? Definitely, no. Seems Nuance and Code Factory have a special relationship, unreachable for others.

-Why is still Nuance selling a discontinued product? Good question, there is no answer. It should be open source, or at least free.

-Is the illegal Eloquence add-on source code legal? Many people believe this code has been written after reverse-engineering the Eloquence included with Jaws. This is not true. The add-on source code was written using the documentation provided in the IBM TTS SDK. This SDK was released as open source by IBM, but became closed after Nuance purchased the rights. So the real question is...

-What happens to a derived work if the original open source software becomes closed? Even reading the IBM TTS Sdk license, we don't know. David, could you please attach this document?

-And the final question: is David's add-on legal? If we answer the previous questions, this one will be answered too. Despite that, David, I think you might request a basic review.

Regards.


El 24/04/2019 a las 7:10, DaVid escribió:
For Patrick: With my IBMTTS driver you can use eloquence or the older
viavoice if you want, but you could use voxin if I implement it from
wsl.
All the three share exactly the same api.

For Shaun: One core and vocalizer voices can't speak as fast as
Eloquence or Espeak.
I use eloquence at 90% and Espeak I don't remember, but I always need
to enable rate boost.
OK maybe vocalizer, but vocalizer is not very understantable at high rates.
Advanced blind users need to be fast in using computers, so we need
fast and understandable synthesizers.
Espeak could be good in english, but is not very good in spanish or
another languages.
I tried to recommend Espeak to new users, but no one liked it.
Bad point for NVDA, and I can't force users to use Espeak. I need to
attract users to this screen reader, so I am looking for a cheap
solution. Vocalizer or eloquence from CF aren't cheap solutions.

Voxin can't be compiled for windows because Oralux doesn't has the source code.
If we want to run voxin with NVDA, we need to use WSL or similar
system to run linux software. I wrote them, they told me that I can
implement a driver for my app, but the final user needs to buy a voxin
copy.
My app = A bridge controller for NVDA, from linux to windows.

I don't know about flite, I will look for a audio demo.

Regards,
David.

2019-04-23 21:15 GMT-06:00, Shaun Everiss <sm.everiss@...>:
Well code factory hold the rights for eloquence so we have to put up
with it.

Then again do you need to use eloquence.

There are the 1 core voices, and vocaliser.

If you really, really need for windows eloquence or a similar, flite
will work, its free, and crappy sounding its used on linux to.

Over that, voxin is only for linux but its cheap enough.

Talk to the devs and see what it would take to make it for nvda and or
sapi.

I mean if its only 30 bucks an os I will buy it gladly.



On 24/04/2019 12:52 PM, DaVid wrote:
Hi all.

I'd like to comment about eloquence and IBMTTS legality. Yes, I'm
reliving a very past commented topic, I know.
I was the first developer of the driver, originally called ibmeci or
eci, I can't remember. At time of developing it, I was a high school
teenager unconcerned of licence problems.

But develop for IBMTTS is not ilegal I think. Voxin and TTSynth let it
for theyr products, but the user must have to buy a personal licence.
E.G. to oralux, it costs just $5 per language.
So the driver IBMTTS should be legal.
Although use it with eloquence libraries would be ilegal and I as a
developer can't control it, it's just life coincidences that eloquence
and IBMTTS share the same api.
OK, voxin or TTSynth doesn't work on windows, but it could be solved
with wsl. I had good results running voxin on wsl, and I could write a
fakade eci library on windows to communicate with voxin in wsl. Oralux
has in their github an eci fakade example.
I just want show more evidence about legality of an IBMTTS driver.
NVDA users could buy voxin and use it or any library with the same
api.

Some weeks ago I loaded a IBMTTS driver to github as a NVDA addon, it
adds many new features and compatibility with python 3, audio ducking
improvements, easier dialog to configure the path of IBM tts library
files, ETC. Also contributions from many of you, but as a ilegal
driver no one wrote his name here.

I'd like to receive contributions here to avoid multiple driver
distributions.
So I want to ask you. If this driver is based on voxin documentation
(And no one can prove otherwise) could it be legal?
Do you consider useful to make voxin compatible with NVDA?

I know about CF driver. But that driver is expensive, outdated, with
many performance problems, requires internet connection, ETC. And CF
doesn't seem to want to solve any of this.
Could be useful to use a IBM tts voice like voxin with NVDA for just $5?

Regards,
David.



.



      
--

José Manuel Delicado Alcolea
Administrador y editor en la web nvda.es
Twitter: @nvda_es
Certificado en el programa NVDA Expert 2019

Logo NVDA Certified Expert


DaVid
 

Hi Jose,
The requirements for wsl are windows 10 1607, or later, and supports
64 bits only.
So wsl shoud run on majority windows 10 distributions. The exception
could be Windows S I think; because "Windows 10 S does not run command
line applications, nor the Windows Console, Cmd / PowerShell, or WSL
instances.
I could write a script to enable wsl feature, and then the user has to
install a linux distribution from the store (E.G. ubuntu) and execute
another script to configure wsl and voxin. In fact I wrote a script to
set voxin on wsl.
that is, the only thing that cannot be automated is the installation
of a linux distribution from the microsoft store.
Speech Dispatcher is not a requirement to run voxin with NVDA. I can
write a bridge on linux to communicate it to windows. So the fakade
IBMTTS library on windows could launch wsl and my voxin server on
linux, I know how to do it without sudo privileges.
there are many Interoperability options to communicate processes from
windows to linux. Using tcp could be an idea but since WSL is not a
virtual machine we could use faster and more direct communication
protocols between processes.
WSL can run without spawning a cmd window.

But Yes, windows 8, 8.1 and 7 are excluded. It could be useful for
working places, E.G. I as a developer can't use eloquence in my
workplace and I don't want to use the expensive and bad implementation
of code factory. So voxin might be a solution.
The NVDA driver doesn't need to be modified for voxin. All relies on
the fakade library.

Here is the complete api reference for IBMTTS. Its fully compatible
with voxin, TTSynth, viavoice, and eloquence.
http://www.wizzardsoftware.com/docs/tts.pdf
Even without this document available, you can optain documentation
from oralux github repo and many useful implementation examples.
Many under gpl, others under bsd licence.

I will consider to ask for a basic review. Depending on how this thread goes.
I'm showing evidence about the legality of my driver.

Regards,
David.

2019-04-24 4:19 GMT-06:00, José Manuel Delicado Alcolea via Groups.Io
<jm.delicado=nvda.es@groups.io>:

Hi David.

Although the WSL approach seems the optimal solution for licensing, I
see many disadvantages for the medium user:

-Making the synthesizer work requires enabling WSL (Windows subsystem
for Linux) support, installing Ubuntu or another distribution from the
Microsoft Store, configuring it, installing Voxin, integrating it with
Speech Dispatcher, configure Speech Dispatcher to listen on a tcp port
and, finally, installing an NVDA add-on.

-WSL works like a container, similar to LXC and Docker. This means that
the only running processes are those you run manually. There are no
daemons, no services unless you start them, so Speech Dispatcher should
be started manually before running NVDA.

-WSL requires Windows 10 Pro / Enterprise / Pro for workstations.
Windows 10 home, 8, 8.1 and 7 are excluded.

In conclussion, only very advanced users could use Eloquence in this
case. However, this story must finish as soon as possible. It has been
opened for many years. We have discussed it in the spanish list many
times, and I'd like to share the results and opened questions here:

-First of all: Code Factory Eloquence is too expensive for users who
only want to use Eloquence in one computer. The add-on is great for
those who will install the license in three computers and download
Vocalizer voices, cheaper than Tiflotecnia's add-on.

-Code Factory add-on loads an additional process in order to work. This
process includes an extra settings interface which allows ignoring and
bypassing the NVDA synth settings ring configuration. In my opinion,
this is a very bad practice, currently NVDA is flexible enough to allow
many configurations for different situations.

-Illegal Eloquence add-on works much better than Code Factory add-on, is
updated more frequently and is compatible with recent NVDA versions. It
has many variants, David is trying to centralize all work in one
repository.

-Illegal Eloquence is written in pure Python, except for Eloquence DLLS,
of course. You can see the entire add-on source code, contribute with
changes, create derived works and so on.

-Also, Tiflotecnia add-on is updated regularly and written in Python. I
recommend purchasing and using it if you want high quality voices.

-First question: why is the illegal Eloquence illegal? We don't have a
clear answer, but We believe the problem comes from the libraries
distribution. Illegal Eloquence has not permission from Nuance to
distribute the DLLS. Would it become legal if libraries are removed? We
don't know.

-Who owns Eloquence currently? Nuance has the ownership, but Code
Factory is the main distributor since 2014.

-Can I be a Code Factory reseller? Yes, but you have to purchase many
pre-paid licenses before selling.

-Can I access the Eloquence SDK and create a legal version derived from
the illegal add-on? Yes, but you have to pay €3500 to Code Factory in
first place, and €18 per sold license.

-Can I become a Nuance distributor? Definitely, no. Seems Nuance and
Code Factory have a special relationship, unreachable for others.

-Why is still Nuance selling a discontinued product? Good question,
there is no answer. It should be open source, or at least free.

-Is the illegal Eloquence add-on source code legal? Many people believe
this code has been written after reverse-engineering the Eloquence
included with Jaws. This is not true. The add-on source code was written
using the documentation provided in the IBM TTS SDK. This SDK was
released as open source by IBM, but became closed after Nuance purchased
the rights. So the real question is...

-What happens to a derived work if the original open source software
becomes closed? Even reading the IBM TTS Sdk license, we don't know.
David, could you please attach this document?

-And the final question: is David's add-on legal? If we answer the
previous questions, this one will be answered too. Despite that, David,
I think you might request a basic review.

Regards.


El 24/04/2019 a las 7:10, DaVid escribió:
For Patrick: With my IBMTTS driver you can use eloquence or the older
viavoice if you want, but you could use voxin if I implement it from
wsl.
All the three share exactly the same api.

For Shaun: One core and vocalizer voices can't speak as fast as
Eloquence or Espeak.
I use eloquence at 90% and Espeak I don't remember, but I always need
to enable rate boost.
OK maybe vocalizer, but vocalizer is not very understantable at high
rates.
Advanced blind users need to be fast in using computers, so we need
fast and understandable synthesizers.
Espeak could be good in english, but is not very good in spanish or
another languages.
I tried to recommend Espeak to new users, but no one liked it.
Bad point for NVDA, and I can't force users to use Espeak. I need to
attract users to this screen reader, so I am looking for a cheap
solution. Vocalizer or eloquence from CF aren't cheap solutions.

Voxin can't be compiled for windows because Oralux doesn't has the source
code.
If we want to run voxin with NVDA, we need to use WSL or similar
system to run linux software. I wrote them, they told me that I can
implement a driver for my app, but the final user needs to buy a voxin
copy.
My app = A bridge controller for NVDA, from linux to windows.

I don't know about flite, I will look for a audio demo.

Regards,
David.

2019-04-23 21:15 GMT-06:00, Shaun Everiss <sm.everiss@gmail.com>:
Well code factory hold the rights for eloquence so we have to put up
with it.

Then again do you need to use eloquence.

There are the 1 core voices, and vocaliser.

If you really, really need for windows eloquence or a similar, flite
will work, its free, and crappy sounding its used on linux to.

Over that, voxin is only for linux but its cheap enough.

Talk to the devs and see what it would take to make it for nvda and or
sapi.

I mean if its only 30 bucks an os I will buy it gladly.



On 24/04/2019 12:52 PM, DaVid wrote:
Hi all.

I'd like to comment about eloquence and IBMTTS legality. Yes, I'm
reliving a very past commented topic, I know.
I was the first developer of the driver, originally called ibmeci or
eci, I can't remember. At time of developing it, I was a high school
teenager unconcerned of licence problems.

But develop for IBMTTS is not ilegal I think. Voxin and TTSynth let it
for theyr products, but the user must have to buy a personal licence.
E.G. to oralux, it costs just $5 per language.
So the driver IBMTTS should be legal.
Although use it with eloquence libraries would be ilegal and I as a
developer can't control it, it's just life coincidences that eloquence
and IBMTTS share the same api.
OK, voxin or TTSynth doesn't work on windows, but it could be solved
with wsl. I had good results running voxin on wsl, and I could write a
fakade eci library on windows to communicate with voxin in wsl. Oralux
has in their github an eci fakade example.
I just want show more evidence about legality of an IBMTTS driver.
NVDA users could buy voxin and use it or any library with the same
api.

Some weeks ago I loaded a IBMTTS driver to github as a NVDA addon, it
adds many new features and compatibility with python 3, audio ducking
improvements, easier dialog to configure the path of IBM tts library
files, ETC. Also contributions from many of you, but as a ilegal
driver no one wrote his name here.

I'd like to receive contributions here to avoid multiple driver
distributions.
So I want to ask you. If this driver is based on voxin documentation
(And no one can prove otherwise) could it be legal?
Do you consider useful to make voxin compatible with NVDA?

I know about CF driver. But that driver is expensive, outdated, with
many performance problems, requires internet connection, ETC. And CF
doesn't seem to want to solve any of this.
Could be useful to use a IBM tts voice like voxin with NVDA for just
$5?

Regards,
David.



.

--

José Manuel Delicado Alcolea
Administrador y editor en la web nvda.es <https://nvda.es>
Twitter: @nvda_es <https://twitter.com/nvda_es>
Certificado en el programa NVDA Expert 2019
<https://certification.nvaccess.org>

Logo NVDA Certified Expert <https://certification.nvaccess.org>




DaVid
 

Klatt synths (Eloquence, decktalk, Orpheus) are currently the best for
fast understandable speaking, small footprint and low latency synths.
Unfortunately companies are not interested in that technology because
it is useful for blind people only.
We need a free open source synth with those features. Not just for
NVDA. For Android, Linux, Mac, raspberry pis, embedded systems, ETC.
I liked Speech Player also.
So (I don't wanted to mention yet but your message requires it hehe) I
am working on develop my own synthesizer from scratch based on speech
player, three more implementations and Klatt papers from 1990.
But is a long term project.
In fact, develop a Klatt synthesizer is not difficult. The complex
part is to get good formant parameters for phonemes. E.G. good
Fricative Consonants are hard to get.
I started research this week. Currently I'm coding a gui interface to
control those parameters from a midi controller and test them easily.
I have a behringer x-touch to test, and anyone with a midi controller
could help.

I can't promise anything right now. I do it only because I like to
learn new things, I don't have so much time to devote exclusively to
this.

Regards,
David.

2019-04-24 3:00 GMT-06:00, David Taylor via Groups.Io
<e.david.taylor=icloud.com@groups.io>:

Hi,


This is it for me too, while it's not my hearing, I find my brain simply
cannot process any other synth at the speed I need, and I too get thrown
by the way the Code Factory addon breaks, which it does, very regularly.
Modern voices, without exception, cause me many, many issues, not least
being speed, terrible pronunciation issues and those awful pauses.
Sadly, it's this simple. If that Eloquence driver breaks, and there is
no replacement that sounds identical, I am forced back to Jaws. I have
to say Speech Player was showing a lot of potential, and I think could
have been made to work, but that development was dropped. I'm sorry to
be so negative, but there literally is no other voice I can deal with on
the computer


Windows One core voices are a great example. On any half decent set of
speakers, something about them jars and gives me a headache, all the
different languages have howling but different pronunciation issues,
they are far to quiet, and even with the latest NVDA update, the
punctuation pauses are far too long for doing any kind of business. Many
of the voices also have too wide and unnatural intonations, making them
very hard to listen to. I could literally use my old Braille 'n Speak
faster than I can use most modern voices, although I doubt that would
even work now we're used to Eloquence. Any replacement must have the
precision of sound and pronunciation, without all the bass that's in
many modern voices which are configured for phones, not for any decent
sound system. I've bought both Acapella and Vocaliser drivers to try as
well, and nothing comes close to Eloquence. The Amy voice from whoever
it is is lovely, but too quiet and again, awful, awful prununciations,
so I can't even do anything quickly with that. So, my choices are, NVDA
19.1 forever, Jaws or somebody makes that addon work.


Cheers

Dave


On 24/04/2019 03:40, Patrick Perdue wrote:
Hi:


As far as I'm concerned, the Code Factory package is essentially
unusable. At best, it's glitchy and unstable, and, yes, that licensing
scheme sure is annoying. I have computers I keep offline for specific
purposes, and that isn't compatible with such operations.

This is why I never bought it.


Honestly, I'm no fan of Eloquence, but it's what I use, because my
hearing has gotten to the point to where I can't really deal with
eSpeak anymore, which I used for several years when I first started
using NVDA.

The one core voices don't do it for me either, as they break in
strange ways, even at slow speed.

So, I'm shamelessly using the not so secret Eloquence driver just so I
can get work done.

I'd probably pay for a personal Voxen license to insure future
compatibility or whatever.


On 4/23/2019 8:52 PM, DaVid wrote:
Hi all.

I'd like to comment about eloquence and IBMTTS legality. Yes, I'm
reliving a very past commented topic, I know.
I was the first developer of the driver, originally called ibmeci or
eci, I can't remember. At time of developing it, I was a high school
teenager unconcerned of licence problems.

But develop for IBMTTS is not ilegal I think. Voxin and TTSynth let it
for theyr products, but the user must have to buy a personal licence.
E.G. to oralux, it costs just $5 per language.
So the driver IBMTTS should be legal.
Although use it with eloquence libraries would be ilegal and I as a
developer can't control it, it's just life coincidences that eloquence
and IBMTTS share the same api.
OK, voxin or TTSynth doesn't work on windows, but it could be solved
with wsl. I had good results running voxin on wsl, and I could write a
fakade eci library on windows to communicate with voxin in wsl. Oralux
has in their github an eci fakade example.
I just want show more evidence about legality of an IBMTTS driver.
NVDA users could buy voxin and use it or any library with the same
api.

Some weeks ago I loaded a IBMTTS driver to github as a NVDA addon, it
adds many new features and compatibility with python 3, audio ducking
improvements, easier dialog to configure the path of IBM tts library
files, ETC. Also contributions from many of you, but as a ilegal
driver no one wrote his name here.

I'd like to receive contributions here to avoid multiple driver
distributions.
So I want to ask you. If this driver is based on voxin documentation
(And no one can prove otherwise) could it be legal?
Do you consider useful to make voxin compatible with NVDA?

I know about CF driver. But that driver is expensive, outdated, with
many performance problems, requires internet connection, ETC. And CF
doesn't seem to want to solve any of this.
Could be useful to use a IBM tts voice like voxin with NVDA for just $5?

Regards,
David.






José Manuel Delicado Alcolea
 

Hi David.

See the lxrun command to automate the distribution installation, it might be helpful.

TCP may be used in virtual machines for earlier Windows versions, and only with testing purposes.

Regarding the provided document: the interesting notes about licensing are on pages 3 and 241 (Apendix A). If your work is derived from any provided sample, you must add an extra copyright header. I don't understand at all the Apendix A.

Regards.


El 24/04/2019 a las 16:30, DaVid escribió:
Hi Jose,
The requirements for wsl are windows 10 1607,  or later, and supports
64 bits only.
So wsl shoud run on majority windows 10 distributions. The exception
could be Windows S I think; because "Windows 10 S does not run command
line applications, nor the Windows Console, Cmd / PowerShell, or WSL
instances.
I could write a script to enable wsl feature, and then the user has to
install a linux distribution from the store (E.G. ubuntu) and execute
another script to configure wsl and voxin. In fact I wrote a script to
set voxin on wsl.
that is, the only thing that cannot be automated is the installation
of a linux distribution from the microsoft store.
Speech Dispatcher is not a requirement to run voxin with NVDA. I can
write a bridge on linux to communicate it to windows. So the fakade
IBMTTS library on windows could launch wsl and my voxin server on
linux, I know how to do it without sudo privileges.
 there are many Interoperability options to communicate processes from
windows to linux. Using tcp could be an idea but since WSL is not a
virtual machine we could use faster and more direct communication
protocols between processes.
WSL can run without spawning a cmd window.

But Yes, windows 8, 8.1 and 7 are excluded. It could be useful for
working places, E.G. I as a developer can't use eloquence in my
workplace and I don't want to use the expensive and bad implementation
of code factory. So voxin might be a solution.
The NVDA driver doesn't need to be modified for voxin. All relies on
the fakade library.

Here is the complete api reference for IBMTTS. Its fully compatible
with voxin, TTSynth, viavoice, and eloquence.
http://www.wizzardsoftware.com/docs/tts.pdf
Even without this document available, you can optain documentation
from oralux github repo and many useful implementation examples.
Many under gpl, others under bsd licence.

I will consider to ask for a basic review. Depending on how this thread goes.
I'm showing evidence about the legality of my driver.

Regards,
David.

2019-04-24 4:19 GMT-06:00, José Manuel Delicado Alcolea via Groups.Io
<jm.delicado@...>:
Hi David.

Although the WSL approach seems the optimal solution for licensing, I
see many disadvantages for the medium user:

-Making the synthesizer work requires enabling WSL (Windows subsystem
for Linux) support, installing Ubuntu or another distribution from the
Microsoft Store, configuring it, installing Voxin, integrating it with
Speech Dispatcher, configure Speech Dispatcher to listen on a tcp port
and, finally, installing an NVDA add-on.

-WSL works like a container, similar to LXC and Docker. This means that
the only running processes are those you run manually. There are no
daemons, no services unless you start them, so Speech Dispatcher should
be started manually before running NVDA.

-WSL requires Windows 10 Pro / Enterprise / Pro for workstations.
Windows 10 home, 8, 8.1 and 7 are excluded.

In conclussion, only very advanced users could use Eloquence in this
case. However, this story must finish as soon as possible. It has been
opened for many years. We have discussed it in the spanish list many
times, and I'd like to share the results and opened questions here:

-First of all: Code Factory Eloquence is too expensive for users who
only want to use Eloquence in one computer. The add-on is great for
those who will install the license in three computers and download
Vocalizer voices, cheaper than Tiflotecnia's add-on.

-Code Factory add-on loads an additional process in order to work. This
process includes an extra settings interface which allows ignoring and
bypassing the NVDA synth settings ring configuration. In my opinion,
this is a very bad practice, currently NVDA is flexible enough to allow
many configurations for different situations.

-Illegal Eloquence add-on works much better than Code Factory add-on, is
updated more frequently and is compatible with recent NVDA versions. It
has many variants, David is trying to centralize all work in one
repository.

-Illegal Eloquence is written in pure Python, except for Eloquence DLLS,
of course. You can see the entire add-on source code, contribute with
changes, create derived works and so on.

-Also, Tiflotecnia add-on is updated regularly and written in Python. I
recommend purchasing and using it if you want high quality voices.

-First question: why is the illegal Eloquence illegal? We don't have a
clear answer, but We believe the problem comes from the libraries
distribution. Illegal Eloquence has not permission from Nuance to
distribute the DLLS. Would it become legal if libraries are removed? We
don't know.

-Who owns Eloquence currently? Nuance has the ownership, but Code
Factory is the main distributor since 2014.

-Can I be a Code Factory reseller? Yes, but you have to purchase many
pre-paid licenses before selling.

-Can I access the Eloquence SDK and create a legal version derived from
the illegal add-on? Yes, but you have to pay €3500 to Code Factory in
first place, and €18 per sold license.

-Can I become a Nuance distributor? Definitely, no. Seems Nuance and
Code Factory have a special relationship, unreachable for others.

-Why is still Nuance selling a discontinued product? Good question,
there is no answer. It should be open source, or at least free.

-Is the illegal Eloquence add-on source code legal? Many people believe
this code has been written after reverse-engineering the Eloquence
included with Jaws. This is not true. The add-on source code was written
using the documentation provided in the IBM TTS SDK. This SDK was
released as open source by IBM, but became closed after Nuance purchased
the rights. So the real question is...

-What happens to a derived work if the original open source software
becomes closed? Even reading the IBM TTS Sdk license, we don't know.
David, could you please attach this document?

-And the final question: is David's add-on legal? If we answer the
previous questions, this one will be answered too. Despite that, David,
I think you might request a basic review.

Regards.


El 24/04/2019 a las 7:10, DaVid escribió:
For Patrick: With my IBMTTS driver you can use eloquence or the older
viavoice if you want, but you could use voxin if I implement it from
wsl.
All the three share exactly the same api.

For Shaun: One core and vocalizer voices can't speak as fast as
Eloquence or Espeak.
I use eloquence at 90% and Espeak I don't remember, but I always need
to enable rate boost.
OK maybe vocalizer, but vocalizer is not very understantable at high
rates.
Advanced blind users need to be fast in using computers, so we need
fast and understandable synthesizers.
Espeak could be good in english, but is not very good in spanish or
another languages.
I tried to recommend Espeak to new users, but no one liked it.
Bad point for NVDA, and I can't force users to use Espeak. I need to
attract users to this screen reader, so I am looking for a cheap
solution. Vocalizer or eloquence from CF aren't cheap solutions.

Voxin can't be compiled for windows because Oralux doesn't has the source
code.
If we want to run voxin with NVDA, we need to use WSL or similar
system to run linux software. I wrote them, they told me that I can
implement a driver for my app, but the final user needs to buy a voxin
copy.
My app = A bridge controller for NVDA, from linux to windows.

I don't know about flite, I will look for a audio demo.

Regards,
David.

2019-04-23 21:15 GMT-06:00, Shaun Everiss <sm.everiss@...>:
Well code factory hold the rights for eloquence so we have to put up
with it.

Then again do you need to use eloquence.

There are the 1 core voices, and vocaliser.

If you really, really need for windows eloquence or a similar, flite
will work, its free, and crappy sounding its used on linux to.

Over that, voxin is only for linux but its cheap enough.

Talk to the devs and see what it would take to make it for nvda and or
sapi.

I mean if its only 30 bucks an os I will buy it gladly.



On 24/04/2019 12:52 PM, DaVid wrote:
Hi all.

I'd like to comment about eloquence and IBMTTS legality. Yes, I'm
reliving a very past commented topic, I know.
I was the first developer of the driver, originally called ibmeci or
eci, I can't remember. At time of developing it, I was a high school
teenager unconcerned of licence problems.

But develop for IBMTTS is not ilegal I think. Voxin and TTSynth let it
for theyr products, but the user must have to buy a personal licence.
E.G. to oralux, it costs just $5 per language.
So the driver IBMTTS should be legal.
Although use it with eloquence libraries would be ilegal and I as a
developer can't control it, it's just life coincidences that eloquence
and IBMTTS share the same api.
OK, voxin or TTSynth doesn't work on windows, but it could be solved
with wsl. I had good results running voxin on wsl, and I could write a
fakade eci library on windows to communicate with voxin in wsl. Oralux
has in their github an eci fakade example.
I just want show more evidence about legality of an IBMTTS driver.
NVDA users could buy voxin and use it or any library with the same
api.

Some weeks ago I loaded a IBMTTS driver to github as a NVDA addon, it
adds many new features and compatibility with python 3, audio ducking
improvements, easier dialog to configure the path of IBM tts library
files, ETC. Also contributions from many of you, but as a ilegal
driver no one wrote his name here.

I'd like to receive contributions here to avoid multiple driver
distributions.
So I want to ask you. If this driver is based on voxin documentation
(And no one can prove otherwise) could it be legal?
Do you consider useful to make voxin compatible with NVDA?

I know about CF driver. But that driver is expensive, outdated, with
many performance problems, requires internet connection, ETC. And CF
doesn't seem to want to solve any of this.
Could be useful to use a IBM tts voice like voxin with NVDA for just
$5?

Regards,
David.



.




--

José Manuel Delicado Alcolea
Administrador y editor en la web nvda.es <https://nvda.es>
Twitter: @nvda_es <https://twitter.com/nvda_es>
Certificado en el programa NVDA Expert 2019
<https://certification.nvaccess.org>

Logo NVDA Certified Expert <https://certification.nvaccess.org>






--

José Manuel Delicado Alcolea
Administrador y editor en la web nvda.es
Twitter: @nvda_es
Certificado en el programa NVDA Expert 2019

Logo NVDA Certified Expert


zvonimir stanečić, 9a5dsz
 

Hi to All,

I should refer to one part of the message.

As for the vocalizer driver,  it is too expensive and it is not updated with the latest Advances including voices and languages.

I am talking about the tiflotecnia driver.

The last version of vocalizer we have is 1.1,

Which unfortunatellz doesn't include Croatian.

 

 

From: nvda-addons@nvda-addons.groups.io <nvda-addons@nvda-addons.groups.io> On Behalf Of Jose Manuel Delicado Alcolea via Groups.Io
Sent: Wednesday, April 24, 2019 12:19 PM
To: nvda-addons@nvda-addons.groups.io
Subject: Re: [nvda-addons] IBMTTS and eloquence driver comments.

 

Hi David.

Although the WSL approach seems the optimal solution for licensing, I see many disadvantages for the medium user:

-Making the synthesizer work requires enabling WSL (Windows subsystem for Linux) support, installing Ubuntu or another distribution from the Microsoft Store, configuring it, installing Voxin, integrating it with Speech Dispatcher, configure Speech Dispatcher to listen on a tcp port and, finally, installing an NVDA add-on.

-WSL works like a container, similar to LXC and Docker. This means that the only running processes are those you run manually. There are no daemons, no services unless you start them, so Speech Dispatcher should be started manually before running NVDA.

-WSL requires Windows 10 Pro / Enterprise / Pro for workstations. Windows 10 home, 8, 8.1 and 7 are excluded.

In conclussion, only very advanced users could use Eloquence in this case. However, this story must finish as soon as possible. It has been opened for many years. We have discussed it in the spanish list many times, and I'd like to share the results and opened questions here:

-First of all: Code Factory Eloquence is too expensive for users who only want to use Eloquence in one computer. The add-on is great for those who will install the license in three computers and download Vocalizer voices, cheaper than Tiflotecnia's add-on.

-Code Factory add-on loads an additional process in order to work. This process includes an extra settings interface which allows ignoring and bypassing the NVDA synth settings ring configuration. In my opinion, this is a very bad practice, currently NVDA is flexible enough to allow many configurations for different situations.

-Illegal Eloquence add-on works much better than Code Factory add-on, is updated more frequently and is compatible with recent NVDA versions. It has many variants, David is trying to centralize all work in one repository.

-Illegal Eloquence is written in pure Python, except for Eloquence DLLS, of course. You can see the entire add-on source code, contribute with changes, create derived works and so on.

-Also, Tiflotecnia add-on is updated regularly and written in Python. I recommend purchasing and using it if you want high quality voices.

-First question: why is the illegal Eloquence illegal? We don't have a clear answer, but We believe the problem comes from the libraries distribution. Illegal Eloquence has not permission from Nuance to distribute the DLLS. Would it become legal if libraries are removed? We don't know.

-Who owns Eloquence currently? Nuance has the ownership, but Code Factory is the main distributor since 2014.

-Can I be a Code Factory reseller? Yes, but you have to purchase many pre-paid licenses before selling.

-Can I access the Eloquence SDK and create a legal version derived from the illegal add-on? Yes, but you have to pay €3500 to Code Factory in first place, and €18 per sold license.

-Can I become a Nuance distributor? Definitely, no. Seems Nuance and Code Factory have a special relationship, unreachable for others.

-Why is still Nuance selling a discontinued product? Good question, there is no answer. It should be open source, or at least free.

-Is the illegal Eloquence add-on source code legal? Many people believe this code has been written after reverse-engineering the Eloquence included with Jaws. This is not true. The add-on source code was written using the documentation provided in the IBM TTS SDK. This SDK was released as open source by IBM, but became closed after Nuance purchased the rights. So the real question is...

-What happens to a derived work if the original open source software becomes closed? Even reading the IBM TTS Sdk license, we don't know. David, could you please attach this document?

-And the final question: is David's add-on legal? If we answer the previous questions, this one will be answered too. Despite that, David, I think you might request a basic review.

Regards.

 

El 24/04/2019 a las 7:10, DaVid escribió:

For Patrick: With my IBMTTS driver you can use eloquence or the older
viavoice if you want, but you could use voxin if I implement it from
wsl.
All the three share exactly the same api.
 
For Shaun: One core and vocalizer voices can't speak as fast as
Eloquence or Espeak.
I use eloquence at 90% and Espeak I don't remember, but I always need
to enable rate boost.
OK maybe vocalizer, but vocalizer is not very understantable at high rates.
Advanced blind users need to be fast in using computers, so we need
fast and understandable synthesizers.
Espeak could be good in english, but is not very good in spanish or
another languages.
I tried to recommend Espeak to new users, but no one liked it.
Bad point for NVDA, and I can't force users to use Espeak. I need to
attract users to this screen reader, so I am looking for a cheap
solution. Vocalizer or eloquence from CF aren't cheap solutions.
 
Voxin can't be compiled for windows because Oralux doesn't has the source code.
If we want to run voxin with NVDA, we need to use WSL or similar
system to run linux software. I wrote them, they told me that I can
implement a driver for my app, but the final user needs to buy a voxin
copy.
My app = A bridge controller for NVDA, from linux to windows.
 
I don't know about flite, I will look for a audio demo.
 
Regards,
David.
 
2019-04-23 21:15 GMT-06:00, Shaun Everiss <sm.everiss@...>:
Well code factory hold the rights for eloquence so we have to put up
with it.
 
Then again do you need to use eloquence.
 
There are the 1 core voices, and vocaliser.
 
If you really, really need for windows eloquence or a similar, flite
will work, its free, and crappy sounding its used on linux to.
 
Over that, voxin is only for linux but its cheap enough.
 
Talk to the devs and see what it would take to make it for nvda and or
sapi.
 
I mean if its only 30 bucks an os I will buy it gladly.
 
 
 
On 24/04/2019 12:52 PM, DaVid wrote:
Hi all.
 
I'd like to comment about eloquence and IBMTTS legality. Yes, I'm
reliving a very past commented topic, I know.
I was the first developer of the driver, originally called ibmeci or
eci, I can't remember. At time of developing it, I was a high school
teenager unconcerned of licence problems.
 
But develop for IBMTTS is not ilegal I think. Voxin and TTSynth let it
for theyr products, but the user must have to buy a personal licence.
E.G. to oralux, it costs just $5 per language.
So the driver IBMTTS should be legal.
Although use it with eloquence libraries would be ilegal and I as a
developer can't control it, it's just life coincidences that eloquence
and IBMTTS share the same api.
OK, voxin or TTSynth doesn't work on windows, but it could be solved
with wsl. I had good results running voxin on wsl, and I could write a
fakade eci library on windows to communicate with voxin in wsl. Oralux
has in their github an eci fakade example.
I just want show more evidence about legality of an IBMTTS driver.
NVDA users could buy voxin and use it or any library with the same
api.
 
Some weeks ago I loaded a IBMTTS driver to github as a NVDA addon, it
adds many new features and compatibility with python 3, audio ducking
improvements, easier dialog to configure the path of IBM tts library
files, ETC. Also contributions from many of you, but as a ilegal
driver no one wrote his name here.
 
I'd like to receive contributions here to avoid multiple driver
distributions.
So I want to ask you. If this driver is based on voxin documentation
(And no one can prove otherwise) could it be legal?
Do you consider useful to make voxin compatible with NVDA?
 
I know about CF driver. But that driver is expensive, outdated, with
many performance problems, requires internet connection, ETC. And CF
doesn't seem to want to solve any of this.
Could be useful to use a IBM tts voice like voxin with NVDA for just $5?
 
Regards,
David.
 
 
 
.
 
 
 
 
 
 
 
 

--

José Manuel Delicado Alcolea
Administrador y editor en la web nvda.es
Twitter: @nvda_es
Certificado en el programa NVDA Expert 2019

Logo NVDA Certified Expert


David Taylor
 

This is very welcome, thank you.


Cheers

Dave

On 24/04/2019 15:35, DaVid wrote:
Klatt synths (Eloquence, decktalk, Orpheus) are currently the best for
fast understandable speaking, small footprint and low latency synths.
Unfortunately companies are not interested in that technology because
it is useful for blind people only.
We need a free open source synth with those features. Not just for
NVDA. For Android, Linux, Mac, raspberry pis, embedded systems, ETC.
I liked Speech Player also.
So (I don't wanted to mention yet but your message requires it hehe) I
am working on develop my own synthesizer from scratch based on speech
player, three more implementations and Klatt papers from 1990.
But is a long term project.
In fact, develop a Klatt synthesizer is not difficult. The complex
part is to get good formant parameters for phonemes. E.G. good
Fricative Consonants are hard to get.
I started research this week. Currently I'm coding a gui interface to
control those parameters from a midi controller and test them easily.
I have a behringer x-touch to test, and anyone with a midi controller
could help.

I can't promise anything right now. I do it only because I like to
learn new things, I don't have so much time to devote exclusively to
this.

Regards,
David.

2019-04-24 3:00 GMT-06:00, David Taylor via Groups.Io
<e.david.taylor=icloud.com@groups.io>:
Hi,


This is it for me too, while it's not my hearing, I find my brain simply
cannot process any other synth at the speed I need, and I too get thrown
by the way the Code Factory addon breaks, which it does, very regularly.
Modern voices, without exception, cause me many, many issues, not least
being speed, terrible pronunciation issues and those awful pauses.
Sadly, it's this simple. If that Eloquence driver breaks, and there is
no replacement that sounds identical, I am forced back to Jaws. I have
to say Speech Player was showing a lot of potential, and I think could
have been made to work, but that development was dropped. I'm sorry to
be so negative, but there literally is no other voice I can deal with on
the computer


Windows One core voices are a great example. On any half decent set of
speakers, something about them jars and gives me a headache, all the
different languages have howling but different pronunciation issues,
they are far to quiet, and even with the latest NVDA update, the
punctuation pauses are far too long for doing any kind of business. Many
of the voices also have too wide and unnatural intonations, making them
very hard to listen to. I could literally use my old Braille 'n Speak
faster than I can use most modern voices, although I doubt that would
even work now we're used to Eloquence. Any replacement must have the
precision of sound and pronunciation, without all the bass that's in
many modern voices which are configured for phones, not for any decent
sound system. I've bought both Acapella and Vocaliser drivers to try as
well, and nothing comes close to Eloquence. The Amy voice from whoever
it is is lovely, but too quiet and again, awful, awful prununciations,
so I can't even do anything quickly with that. So, my choices are, NVDA
19.1 forever, Jaws or somebody makes that addon work.


Cheers

Dave


On 24/04/2019 03:40, Patrick Perdue wrote:
Hi:


As far as I'm concerned, the Code Factory package is essentially
unusable. At best, it's glitchy and unstable, and, yes, that licensing
scheme sure is annoying. I have computers I keep offline for specific
purposes, and that isn't compatible with such operations.

This is why I never bought it.


Honestly, I'm no fan of Eloquence, but it's what I use, because my
hearing has gotten to the point to where I can't really deal with
eSpeak anymore, which I used for several years when I first started
using NVDA.

The one core voices don't do it for me either, as they break in
strange ways, even at slow speed.

So, I'm shamelessly using the not so secret Eloquence driver just so I
can get work done.

I'd probably pay for a personal Voxen license to insure future
compatibility or whatever.


On 4/23/2019 8:52 PM, DaVid wrote:
Hi all.

I'd like to comment about eloquence and IBMTTS legality. Yes, I'm
reliving a very past commented topic, I know.
I was the first developer of the driver, originally called ibmeci or
eci, I can't remember. At time of developing it, I was a high school
teenager unconcerned of licence problems.

But develop for IBMTTS is not ilegal I think. Voxin and TTSynth let it
for theyr products, but the user must have to buy a personal licence.
E.G. to oralux, it costs just $5 per language.
So the driver IBMTTS should be legal.
Although use it with eloquence libraries would be ilegal and I as a
developer can't control it, it's just life coincidences that eloquence
and IBMTTS share the same api.
OK, voxin or TTSynth doesn't work on windows, but it could be solved
with wsl. I had good results running voxin on wsl, and I could write a
fakade eci library on windows to communicate with voxin in wsl. Oralux
has in their github an eci fakade example.
I just want show more evidence about legality of an IBMTTS driver.
NVDA users could buy voxin and use it or any library with the same
api.

Some weeks ago I loaded a IBMTTS driver to github as a NVDA addon, it
adds many new features and compatibility with python 3, audio ducking
improvements, easier dialog to configure the path of IBM tts library
files, ETC. Also contributions from many of you, but as a ilegal
driver no one wrote his name here.

I'd like to receive contributions here to avoid multiple driver
distributions.
So I want to ask you. If this driver is based on voxin documentation
(And no one can prove otherwise) could it be legal?
Do you consider useful to make voxin compatible with NVDA?

I know about CF driver. But that driver is expensive, outdated, with
many performance problems, requires internet connection, ETC. And CF
doesn't seem to want to solve any of this.
Could be useful to use a IBM tts voice like voxin with NVDA for just $5?

Regards,
David.




James Scholes
 

I would love to pay for a legal Eloquence driver that wasn't terrible, but the only person who can answer legal questions is a lawyer. We can't give you legal advice on this list, sorry.

Regards,

James Scholes

On 23/04/2019 at 7:52 pm, DaVid wrote:
Hi all.
I'd like to comment about eloquence and IBMTTS legality. Yes, I'm
reliving a very past commented topic, I know.
I was the first developer of the driver, originally called ibmeci or
eci, I can't remember. At time of developing it, I was a high school
teenager unconcerned of licence problems.
But develop for IBMTTS is not ilegal I think. Voxin and TTSynth let it
for theyr products, but the user must have to buy a personal licence.
E.G. to oralux, it costs just $5 per language.
So the driver IBMTTS should be legal.
Although use it with eloquence libraries would be ilegal and I as a
developer can't control it, it's just life coincidences that eloquence
and IBMTTS share the same api.
OK, voxin or TTSynth doesn't work on windows, but it could be solved
with wsl. I had good results running voxin on wsl, and I could write a
fakade eci library on windows to communicate with voxin in wsl. Oralux
has in their github an eci fakade example.
I just want show more evidence about legality of an IBMTTS driver.
NVDA users could buy voxin and use it or any library with the same
api.
Some weeks ago I loaded a IBMTTS driver to github as a NVDA addon, it
adds many new features and compatibility with python 3, audio ducking
improvements, easier dialog to configure the path of IBM tts library
files, ETC. Also contributions from many of you, but as a ilegal
driver no one wrote his name here.
I'd like to receive contributions here to avoid multiple driver distributions.
So I want to ask you. If this driver is based on voxin documentation
(And no one can prove otherwise) could it be legal?
Do you consider useful to make voxin compatible with NVDA?
I know about CF driver. But that driver is expensive, outdated, with
many performance problems, requires internet connection, ETC. And CF
doesn't seem to want to solve any of this.
Could be useful to use a IBM tts voice like voxin with NVDA for just $5?
Regards,
David.


James Scholes
 

José Manuel Delicado Alcolea wrote:
-WSL requires Windows 10 Pro / Enterprise / Pro for workstations.
Windows 10 home, 8, 8.1 and 7 are excluded.
WSL works on Windows 10 Home, but agreed on the rest of it.

Regards,

James Scholes

On 24/04/2019 at 5:19 am, José Manuel Delicado Alcolea via Groups.Io wrote:
Hi David.
Although the WSL approach seems the optimal solution for licensing, I see many disadvantages for the medium user:
-Making the synthesizer work requires enabling WSL (Windows subsystem for Linux) support, installing Ubuntu or another distribution from the Microsoft Store, configuring it, installing Voxin, integrating it with Speech Dispatcher, configure Speech Dispatcher to listen on a tcp port and, finally, installing an NVDA add-on.
-WSL works like a container, similar to LXC and Docker. This means that the only running processes are those you run manually. There are no daemons, no services unless you start them, so Speech Dispatcher should be started manually before running NVDA.
-WSL requires Windows 10 Pro / Enterprise / Pro for workstations. Windows 10 home, 8, 8.1 and 7 are excluded.
In conclussion, only very advanced users could use Eloquence in this case. However, this story must finish as soon as possible. It has been opened for many years. We have discussed it in the spanish list many times, and I'd like to share the results and opened questions here:
-First of all: Code Factory Eloquence is too expensive for users who only want to use Eloquence in one computer. The add-on is great for those who will install the license in three computers and download Vocalizer voices, cheaper than Tiflotecnia's add-on.
-Code Factory add-on loads an additional process in order to work. This process includes an extra settings interface which allows ignoring and bypassing the NVDA synth settings ring configuration. In my opinion, this is a very bad practice, currently NVDA is flexible enough to allow many configurations for different situations.
-Illegal Eloquence add-on works much better than Code Factory add-on, is updated more frequently and is compatible with recent NVDA versions. It has many variants, David is trying to centralize all work in one repository.
-Illegal Eloquence is written in pure Python, except for Eloquence DLLS, of course. You can see the entire add-on source code, contribute with changes, create derived works and so on.
-Also, Tiflotecnia add-on is updated regularly and written in Python. I recommend purchasing and using it if you want high quality voices.
-First question: why is the illegal Eloquence illegal? We don't have a clear answer, but We believe the problem comes from the libraries distribution. Illegal Eloquence has not permission from Nuance to distribute the DLLS. Would it become legal if libraries are removed? We don't know.
-Who owns Eloquence currently? Nuance has the ownership, but Code Factory is the main distributor since 2014.
-Can I be a Code Factory reseller? Yes, but you have to purchase many pre-paid licenses before selling.
-Can I access the Eloquence SDK and create a legal version derived from the illegal add-on? Yes, but you have to pay €3500 to Code Factory in first place, and €18 per sold license.
-Can I become a Nuance distributor? Definitely, no. Seems Nuance and Code Factory have a special relationship, unreachable for others.
-Why is still Nuance selling a discontinued product? Good question, there is no answer. It should be open source, or at least free.
-Is the illegal Eloquence add-on source code legal? Many people believe this code has been written after reverse-engineering the Eloquence included with Jaws. This is not true. The add-on source code was written using the documentation provided in the IBM TTS SDK. This SDK was released as open source by IBM, but became closed after Nuance purchased the rights. So the real question is...
-What happens to a derived work if the original open source software becomes closed? Even reading the IBM TTS Sdk license, we don't know. David, could you please attach this document?
-And the final question: is David's add-on legal? If we answer the previous questions, this one will be answered too. Despite that, David, I think you might request a basic review.
Regards.
El 24/04/2019 a las 7:10, DaVid escribió:
For Patrick: With my IBMTTS driver you can use eloquence or the older
viavoice if you want, but you could use voxin if I implement it from
wsl.
All the three share exactly the same api.

For Shaun: One core and vocalizer voices can't speak as fast as
Eloquence or Espeak.
I use eloquence at 90% and Espeak I don't remember, but I always need
to enable rate boost.
OK maybe vocalizer, but vocalizer is not very understantable at high rates.
Advanced blind users need to be fast in using computers, so we need
fast and understandable synthesizers.
Espeak could be good in english, but is not very good in spanish or
another languages.
I tried to recommend Espeak to new users, but no one liked it.
Bad point for NVDA, and I can't force users to use Espeak. I need to
attract users to this screen reader, so I am looking for a cheap
solution. Vocalizer or eloquence from CF aren't cheap solutions.

Voxin can't be compiled for windows because Oralux doesn't has the source code.
If we want to run voxin with NVDA, we need to use WSL or similar
system to run linux software. I wrote them, they told me that I can
implement a driver for my app, but the final user needs to buy a voxin
copy.
My app = A bridge controller for NVDA, from linux to windows.

I don't know about flite, I will look for a audio demo.

Regards,
David.

2019-04-23 21:15 GMT-06:00, Shaun Everiss<sm.everiss@gmail.com>:
Well code factory hold the rights for eloquence so we have to put up
with it.

Then again do you need to use eloquence.

There are the 1 core voices, and vocaliser.

If you really, really need for windows eloquence or a similar, flite
will work, its free, and crappy sounding its used on linux to.

Over that, voxin is only for linux but its cheap enough.

Talk to the devs and see what it would take to make it for nvda and or
sapi.

I mean if its only 30 bucks an os I will buy it gladly.



On 24/04/2019 12:52 PM, DaVid wrote:
Hi all.

I'd like to comment about eloquence and IBMTTS legality. Yes, I'm
reliving a very past commented topic, I know.
I was the first developer of the driver, originally called ibmeci or
eci, I can't remember. At time of developing it, I was a high school
teenager unconcerned of licence problems.

But develop for IBMTTS is not ilegal I think. Voxin and TTSynth let it
for theyr products, but the user must have to buy a personal licence.
E.G. to oralux, it costs just $5 per language.
So the driver IBMTTS should be legal.
Although use it with eloquence libraries would be ilegal and I as a
developer can't control it, it's just life coincidences that eloquence
and IBMTTS share the same api.
OK, voxin or TTSynth doesn't work on windows, but it could be solved
with wsl. I had good results running voxin on wsl, and I could write a
fakade eci library on windows to communicate with voxin in wsl. Oralux
has in their github an eci fakade example.
I just want show more evidence about legality of an IBMTTS driver.
NVDA users could buy voxin and use it or any library with the same
api.

Some weeks ago I loaded a IBMTTS driver to github as a NVDA addon, it
adds many new features and compatibility with python 3, audio ducking
improvements, easier dialog to configure the path of IBM tts library
files, ETC. Also contributions from many of you, but as a ilegal
driver no one wrote his name here.

I'd like to receive contributions here to avoid multiple driver
distributions.
So I want to ask you. If this driver is based on voxin documentation
(And no one can prove otherwise) could it be legal?
Do you consider useful to make voxin compatible with NVDA?

I know about CF driver. But that driver is expensive, outdated, with
many performance problems, requires internet connection, ETC. And CF
doesn't seem to want to solve any of this.
Could be useful to use a IBM tts voice like voxin with NVDA for just $5?

Regards,
David.



.
--
José Manuel Delicado Alcolea
Administrador y editor en la web nvda.es <https://nvda.es>
Twitter: @nvda_es <https://twitter.com/nvda_es>
Certificado en el programa NVDA Expert 2019 <https://certification.nvaccess.org>
Logo NVDA Certified Expert <https://certification.nvaccess.org>


Noelia Ruiz
 

I'm not interested in Eloquence.
Anyway, as a reviewer, if you request for basic review, considering that in an old blog post by an NV Access Developer was say that, at that moment, using Eloquence with NVDA was illegal, I would ask NV Access in case they can give any answer.
In doubt, I wouldn't continue with basic review and I think we need precisse answers for this.
Unfortunately now I can't find the mentioned post, referenced in the Eoquence section of this document:
https://github.com/nvaccess/nvda-community/wiki/SwitchingFromJawsToNVDA

Eloquence
One of the most asked questions concerns the use of the Eloquence synthesizer with NVDA. Until recently, *it was illegal to do so, as explained by a developer.*
The broken link points to
http://community.nvda-project.org/blog/NVDAAndEloquenceSituation

Cheers

El 24/04/2019 a las 19:56, James Scholes escribió:
I would love to pay for a legal Eloquence driver that wasn't terrible, but the only person who can answer legal questions is a lawyer.  We can't give you legal advice on this list, sorry.
Regards,
James Scholes
On 23/04/2019 at 7:52 pm, DaVid wrote:
Hi all.

I'd like to comment about eloquence and IBMTTS legality. Yes, I'm
reliving a very past commented topic, I know.
I was the first developer of the driver, originally called ibmeci or
eci, I can't remember. At time of developing it, I was a high school
teenager unconcerned of licence problems.

But develop for IBMTTS is not ilegal I think. Voxin and TTSynth let it
for theyr products, but the user must have to buy a personal licence.
E.G. to oralux, it costs just $5 per language.
So the driver IBMTTS should be legal.
Although use it with eloquence libraries would be ilegal and I as a
developer can't control it, it's just life coincidences that eloquence
and IBMTTS share the same api.
OK, voxin or TTSynth doesn't work on windows, but it could be solved
with wsl. I had good results running voxin on wsl, and I could write a
fakade eci library on windows to communicate with voxin in wsl. Oralux
has in their github an eci fakade example.
I just want show more evidence about legality of an IBMTTS driver.
NVDA users could buy voxin and use it or any library with the same
api.

Some weeks ago I loaded a IBMTTS driver to github as a NVDA addon, it
adds many new features and compatibility with python 3, audio ducking
improvements, easier dialog to configure the path of IBM tts library
files, ETC. Also contributions from many of you, but as a ilegal
driver no one wrote his name here.

I'd like to receive contributions here to avoid multiple driver distributions.
So I want to ask you. If this driver is based on voxin documentation
(And no one can prove otherwise) could it be legal?
Do you consider useful to make voxin compatible with NVDA?

I know about CF driver. But that driver is expensive, outdated, with
many performance problems, requires internet connection, ETC. And CF
doesn't seem to want to solve any of this.
Could be useful to use a IBM tts voice like voxin with NVDA for just $5?

Regards,
David.



Timothy
 

On Wednesday, April 24, 2019 2:51 PM, "Noelia Ruiz" <nrm1977@gmail.com> wrote:
Unfortunately now I can't find the mentioned post, referenced in the
Eoquence section of this document:
https://github.com/nvaccess/nvda-community/wiki/SwitchingFromJawsToNVDA

Eloquence
One of the most asked questions concerns the use of the Eloquence
synthesizer with NVDA. Until recently, *it was illegal to do so, as
explained by a developer.*
The broken link points to
http://community.nvda-project.org/blog/NVDAAndEloquenceSituation
As NV Access does not seem to have a search facility, I went ahead and found the aforementioned blog post on the Wayback Machine:
http://web.archive.org/web/20150923083641/http://community.nvda-project.org/blog/NVDAAndEloquenceSituation

Timothy


José Manuel Delicado Alcolea
 

Thank you very much, Timothy. I read this article in 2010, and after nvda-project website changed to nvaccess.org I searched it with no luck.

Although I don't like it, seems this article gives answers to all previous questions. I will re-read it multiple times to see if something has changed since 2010-2013.

Regards.


El 24/04/2019 a las 21:05, Timothy via Groups.Io escribió:
On Wednesday, April 24, 2019 2:51 PM, "Noelia Ruiz" <nrm1977@...> wrote:
Unfortunately now I can't find the mentioned post, referenced in the 
Eoquence section of this document:
https://github.com/nvaccess/nvda-community/wiki/SwitchingFromJawsToNVDA

Eloquence
One of the most asked questions concerns the use of the Eloquence 
synthesizer with NVDA. Until recently, *it was illegal to do so, as 
explained by a developer.*
The broken link points to
http://community.nvda-project.org/blog/NVDAAndEloquenceSituation
As NV Access does not seem to have a search facility, I went ahead and found the aforementioned blog post on the Wayback Machine:
http://web.archive.org/web/20150923083641/http://community.nvda-project.org/blog/NVDAAndEloquenceSituation

Timothy



--

José Manuel Delicado Alcolea
Administrador y editor en la web nvda.es
Twitter: @nvda_es
Certificado en el programa NVDA Expert 2019

Logo NVDA Certified Expert


DaVid
 

The add-on doesn't reffer to eloquence, does not include eloquence
libraries, but is based on a public document reference api, also on
some oralux examples.
If this add-on can be made for voxin, it should be legal, I think.
because is not eloquence. Oralux let developers to use voxin libraries
as long as the user buys a license and the developer can't include
voxin libraries. Its the user's responsibility to buy and use voxin
over wsl with NVDA or use any other illegal library as many currently
do.
So distribution of an IBMTTS add-on (without libraries) could be legal
as speech dispatcher or emacspeak (for example) are.
I'm looking for solutions. My country is considering to buy a global
country licence of a very expencive screen reader (that I don't want
to mention) and users comments can influence the decision.
Even if my main synthesizer were Espeak, I would try to find a good
and acceptable solution for everyone.

Regards,
David.


James Scholes
 

I agree with you, but all I can do is repeat: if you want to know if distributing certain code is legal or not, you will need to consult with a lawyer. Nobody on this list is qualified or contracted to say that it is legal or otherwise, and anybody giving you advice here won't be able to help you if that advice turns out to be wrong.

Regards,

James Scholes

On 24/04/2019 at 3:10 pm, DaVid wrote:
The add-on doesn't reffer to eloquence, does not include eloquence
libraries, but is based on a public document reference api, also on
some oralux examples.
If this add-on can be made for voxin, it should be legal, I think.
because is not eloquence. Oralux let developers to use voxin libraries
as long as the user buys a license and the developer can't include
voxin libraries. Its the user's responsibility to buy and use voxin
over wsl with NVDA or use any other illegal library as many currently
do.
So distribution of an IBMTTS add-on (without libraries) could be legal
as speech dispatcher or emacspeak (for example) are.
I'm looking for solutions. My country is considering to buy a global
country licence of a very expencive screen reader (that I don't want
to mention) and users comments can influence the decision.
Even if my main synthesizer were Espeak, I would try to find a good
and acceptable solution for everyone.
Regards,
David.